The Empirical Evidence on the Disposition Effect Finance Essay

Disposition effect, which is studied by many literatures, is one of popular bias of investors. It was firstly suggested by Shefrin and Statman (1985). They gave an explanation of the disposition effect that “Investors have tendency to sell winners too early and ride losers too long.” Moreover, the result from the experiment of Odean (1998) supported the disposition effect theory. Odean (1998) compared between Proportion of Gains Realized (PGR) and Proportion of Losses Realized (PLR). PGR can be calculated by dividing realized gains by the sum of realized gains and paper gains and PLR can be calculated by dividing realized losses by the sum of realized losses and paper losses. Realized gain and loss is the gain or loss amount calculated by the difference of sell price and purchasing price when investors sell stocks. Paper gain and loss is an gain and loss amount calculated by the difference between current price and purchasing price for stocks which had not been sold yet.

The data was obtained from entire year trade provided by a worldwide discount brokerage house in the period from 1987 to 1990 and from 1991 to 1993 to calculate PGR and PLR. The result showed that PGR values were greater than PLR values in both periods. It means that investors had tendency to realize their gains more than their losses and this was consistent with the disposition effect. Shefrin and Statman (1985) also pointed out about the theories that were possible causes of the disposition effect which are 1) the prospect theory 2) the theory of seeking pride and avoiding regret 3) the theory of mental accounting and 4) the theory of mean reversion. After that, there were many studies discussed whether those theories can be causes of the disposition effect. Next of this report will discuss about those theories and will talk about some interesting factors which affect the disposition effect.

Body

Start with prospect theory which is the most popular theory which is used to explain the disposition effect, prospect theory was suggested by Kahneman and Tversky (1979). They explained that the prospect theory can be distinguished into two phases which are editing and evaluating. In editing phase, investors compare their choices with the reference point which normally related to their current asset position. The choices which are less than the reference point will be regard as losses and the choices which are greater than the reference point will be regarded as gains. Then, move to evaluating phase, s- shaped valuation function show utility value in vertical line and amount of gain and loss in horizontal line and is concave in gain area and convex in loss area. It will be used for decision making.

Investors always choose a choice which gives them the highest utility. Another important point in the prospect theory is losses hurt more than gain. Therefore, investors will be risk seeker in loss area because they want to avoid the losses but they will be risk averse in gain area because they want to ensure that they will exactly get gains. Next, the relationship between the disposition effect and the prospect theory will be explained. Shefrin and Statman (1985) pointed out that prospect theory can be used to explain the disposition effect because when the stocks gain, the investor who act as risk averse will sell those stocks because they want to get gains surely. On the other hands, when the stocks loss, the investors who act as risk seeker, will keep those stocks because they decide to take risk that the price of the stocks could either rise until investors can get the profit or continuously fall that investors get more losses. This characteristic of the prospect theory is consistent with the disposition effect. Weber and Camerer (1998) conducted experiment using purchasing price as a reference point from both last-in, first-out method (LIFO), which is the method that investors will sell the stock which is bought last, and first-in, first-out method (FIFO), which is the method that investors will sell the stock which is bough first, to compare to selling price, the result showed that the percentage to sell the winning stock is higher than losing stock in both methods.

They used the idea that prospect theory can explain the disposition effect to explain the result of their experiment “In the prospect theory, disposition effects occur because subjects use their purchase price and are reluctant to recognize losses; they gamble in the domain of losses and avoid risk in the domain of gain”. However, there are some comments which show that the prospect theory cannot totally explain the disposition effect. Start with Barberis and Xiong (2009), they pointed out about the case of investor who has prospect theory using annual gain/loss method that “investors will be more inclined to sell stocks with prior losses than stocks with prior gains.” Investors using annual gain/loss means that investor counted their profit by calculating the price of stock between purchasing price and the price at the reference day such as a half year or a one year later from buying those stocks.

Therefore, during the period, it is possible for investors to buy more stocks if they think that stocks will continue to rise in the future or sell them out if they think stocks will continue to fall in the future to accomplish their goal. In their experiment, they divided investment period into three periods which are T0, T1 and T2. T0 is the date investors bought stocks. They showed that even though the stock price fall in time T1 but if they rise in time T2, the overall profit compare between time T2 and time T0 could be small profit. In the same manner, if stocks rise in time T1 and then fall in time T2, investors would get small profit calculated from time T2 and T0. They claimed that since investors were risk averse, they would buy stocks at time 0 if the expected return of stocks were high enough. Therefore, if investors gained in time t1, Baberis and Xiong showed the result from their formula that investor had to buy more stocks for increasing in share allocation to take large gamble that those stocks would rise to get more profit in time T2 than they do so after the stock fall in time t 1 to gamble those stock price would rise and they would get small profit in time T2. Therefore, it contradicts with the characteristic of disposition effect in which investors have a more propensity to sell stocks when stocks rise more than in case that stocks fall. Furthermore, they also conducted an experiment that investigated the share allocation of investors after the stock price rose and fell comparing between using annual gain and loss and realized gain and loss.

Realized gain and loss can be calculated each time when stocks price move. The result showed that using realized gain and loss follow the disposition effect more than annual gain and loss. They gave explanation that using annual gain and loss, investors are forced to accumulate the profit in a single amount in the specific reference time but investors can split to realize their profit in each time the stock price move in realized gain and loss. In case losses, investors do not want to separate their losses in many components while they want to do in case of gain. Therefore, they sell stocks when stocks rise and keep stocks when they fall which was consistent with the disposition effect. Kaustia (2010) raised some different aspect between the disposition effect and the prospect theory. He pointed out that the propensity to sell in the disposition effect is increase or constant when the gain of stocks increase and is constant or insensitive in case of the increase of losses of stock while the prospect theory value function explained that if the gain of the stock price increase, it will minimize the propensity to sell stocks and if the losses of stocks increase, it will decrease to propensity to sell. Seeking pride and Avoiding regret is one theory that explains the disposition effect. Shefrin and Statman (1985) argued that investors held their losing stocks because they felt regret to realize the losses and they felt pride if they could close the stock account with a gain.

Baraber and Odean (1999) illustrated that since the investor avoided regret, this bias would render investors to sell those stocks when the price of stocks rose because they were afraid that the price of the stocks would fall in the future. On the contrary, if the prices of the stocks they held fell, they would hold it because they expected that the price would rise in the future Third, the disposition effect can be explained by the mental accounting theory. Kahneman and Lovallo (1993) pointed out that investors did not concern in overall profit. All outcomes of an individual investor would be separate into several parts. Thaler (1999) said that investors would get more hurt from realized losses than from paper losses. It was painful to close their accounts with losses.

Such an incident rendered investor to hesitate to sell stocks of which the price fell. Therefore, mental theory explains the bias of investors in the same way of the disposition effect does. Another causes of disposition effect comes from mean reversion in which Lakonishok and Smidt (1986) as cited in Odean (1998) explained that the investors would sell the winning stocks because they believed that those stocks had reflected the information already and would keep losing stocks because they believed that the information had not incorporated into price of those stocks yet but it would be in the future. There was an experiment to support mean reversion idea from Weber and Camerer (1998). They investigated the reaction of investors after holding the stock two periods. They found that investors had more percentage to sell the stocks if the price of stock at trading date rose in the second period even though it fell in the first period than those stocks of which their price at trading date fell in the second period even though it rose in the first period. However, what investors expect from mean reversion might not be true in reality. To prove this, Odean (1998) conducted the experiment that followed the excess return next 84 days, 252 days (1 year) and 504 days (2 years) after the last trading day of both the winning stocks sold and paper loss stocks.

The result clearly showed that the excess return of winning stocks was still positive in all three periods and the excess return of paper loss stocks were still negative in all three periods. Therefore, the result showed us that investors cannot be sure that they would get profit from acting follow mean reversion. However, Kausita (2010) did not agree the idea that the theory of mean reversion can use to explain the disposition effect.

She pointed out that investors who followed mean reversion would sell stocks which had recently risen even though they were paper losses and they would keep stocks which had recently fallen even though they were paper gain. The way investors follow was opposite with the disposition effect. Apart from those theories concern above, there are some factors that affect the disposition effect. The first factor is tax concern which always lessens the power of disposition effect in December. There are many literatures explaining about this topic with various reasons. Shefrin and Statman (1985) pointed out the behavior of investors to realize the loss will happen in December more often than other months as a self control strategy.

They explained that investors will normally ride the losing stock but they will have self motivation to realize losses heavily in December which is the deadline of a year to accomplish their tax plan. Odean (1998) also studied about tax effect. He conducted experiment by comparing PGR and PLR by using data in December from 1987 to 1990 and from 1991 to 1993. The result showed that PGR was smaller than PLR in both periods which was the opposite the result when calculating from entire year that PGR was greater than PLR. Odean (1998) wanted to ensure that the difference result came from only tax concern factor, hence, he eliminated rebalancing factor by not including the stocks that sold partially and the stocks repurchased after sold in the experiment, the result still showed that PLR in December was higher than PGR. Therefore, his result stressed the point that investors increased realizing the loss or, in other words, the disposition effect less affect in December because of tax concern not rebalancing issue. Demographic is the second factor that affects disposition effect. Dhar and Zhu (2006) showed that the disposition effect can be affected by the level of income and occupation of investors. They conducted experiment by comparing the disposition effect value among different income level of investors and between investors who work in professional occupation and those who work in non professional occupation. The result showed that high income investors would have lower disposition effect than those with middle income level and middle income investors would have lower disposition effect than those with low income. Furthermore, investors who work in professional occupations would have smaller disposition effect than those who work in non professional occupation.

They explained the result that the high income investors had potential to hire financial advisory firm to help them to minimize bias in their investments and investors who work in professional firms presumably had high financial knowledge which led them to minimize their investment bias. They also conducted the same experiment but used information only in December. Disposition effect value in each group in December was smaller than the value calculated entire year which was consistent with Odean (1998) and the result also showed in the same manner with the above experiment. From the result, it showed that high income investors and those who work in professional occupations would be possibly more concern about tax than those who have middle or low income and work in non professional occupation. Besides, there is evidence show that the house money effect surprisingly coexist the disposition effect even though both of them affect investors in the opposite way.

Thaler and Johnson(1990) as cited in Duxbury et al (2011) explained that the house money effect was a bias that investors had a propensity to take more risk after prior gain of their stocks because they included their prior outcomes into current making decision. In other words, investors could be more tolerate to take risk if they got prior gain but they would be more risk averse if they got prior losses. There are many studies showing that the house money effect existed in many stock markets such as in Australia, Taiwan and Sweden. Duxbury et al (2011) also supported the issue that the disposition effect and the house money effect coexisted. They compared between PGR and PLR by using data from Chinese stock market from February 2001 to December 2004. The result showed that PGR was greater than PLR if all stocks from raw data were observed. Such an incident means that investors had propensity to sell when stocks gained and kept stocks when they lose.

Thus, it was consistent with the disposition effect. They also observed PGR and PLR in case of stocks with prior realized gains and stocks with prior realized losses from the same raw data.

The result showed that PGR and PLR in stocks with prior realized gains will be smaller than those in stocks with realized losses. It means that investors hold more stocks after stocks had prior gains than after stocks had prior losses. When investors held stocks, they took risk that the price of stocks would fall in the future. It can be said that investors were more tolerate to take risk after their stocks gain. This was consistent with the characteristic of the house money effect. Therefore, we can find both the disposition effect and the hose money effect in the same raw data but in different point of view.

They also conducted experiment to find the correlation between the two effects. They found that there was 53.5% from the sample population had both effect simultaneously and the experiment also showed that the correlation between magnitudes of two effects in group of investors who had two effects simultaneously was highly significant and negative. Duxbury et al (2011) explained that the power of the disposition effect would be weaker because of the presence of the house money effect. In other words, when investors were affected by the house money effect which leaded them to take more risk because of prior realized gain in stocks, they would be less affected by the disposition effect. When investors experienced prior realized losses in stocks, they would be less affected by the house money effect and would be more affected by disposition effect which leads them to be more risk averse.

Conclusion

In conclusion, the disposition effect which was suggested by Shefrin and Statman (1985) is one of a popular bias of investors. It explains that investors will have tendency to sell stocks if they gain but tend to keep stocks if they loss. Many theories are brought to explain the disposition effect. They are the prospect theory, the theory of seeking pride and avoiding regret, the theory of mental accounting and the theory of mean reversion.

Many literatures compare the characteristic of those theories and the disposition effect. Some of them support that those theories can explain the disposition effect. Others show criteria which leads some of those theories to have different characteristic from the disposition effect. Moreover, there are some factors which can affect the disposition effect. Those are tax concern and demographic.

Besides, there is evidence from Duxbury et al (2011) showed that the house money effect and the disposition effect surprisingly coexist even though both of them affect investors in the opposite way.

Cite this page